Research On Technologies Of Infrared Target Tracking And Distance Sensing

Infrared detection is a kind of passive detection,which has the characteristics of high concealment and strong anti-interference etc.Infrared imaging technology has a wide range of applications in civil and national defense.Infrared target tracking and distance sensing technology are two important research directions of infrared image application.Infrared target tracking and distance sensing technology can be widely used in the fields of infrared imaging guidance,monitoring and warning.Therefore,it has important theoretical and practical value to deeply research the target tracking algorithms under a complex background and target distance sensing.In this paper,based on the existing visible and infrared target tracking technology,in order to increase the difference between the target and the background and extract the effective target features,the curvature filter and guided filter are introduced to preprocess the image patch.Three infrared target tracking algorithms are analyzed and verified.At the same time,using the classical light scattering model and meteorological parameter,a distance sensing algorithm for urban buildings in a foggy day is introduced.The main research results are as follows.(1)The performance evaluation of infrared dim small target tracking algorithm requires a large number of test images with known positions and parameters of infrared dim and small target.Aiming at the problem that it is impossible to obtain a large number of infrared dim and small target images with continuous changes in parameters in a short time,this paper proposed a simulation method of infrared dim small target image sequence.A large number of infrared dim and small target images can be generated using the proposed method,and the parameters of targets could be controlled.In this paper,the imaging characteristics of infrared dim small target on the image is analyzed deeply.Then the mathematical simulation model of infrared camera shake is built here.Finally,a infrared dim-small target image sequence simulation method is developed.(2)Aiming at the problem that infrared dim and small target tracking is easy to occur offset under complex sky background.In this paper,the curvature filter is used to process the spatio-temporal context prior model.And then the difference image of the original input image and the filtering result is sent to the spatio-temporal context model as a new spatio-temporal context prior.An infrared dim-small target tracking algorithm combining spatiotemporal context and curvature filtering is proposed in this paper.(3)In order to solve the problem of tracking failure caused by maneuvering in the process of infrared dim and small target moving,on the basis of kernelized correlation filter,a guided filter is introduced here to process the base samples in order to increase the discrimination between the infrared dim target positive sample and the background negative sample.The difference image between the original image and the filtering result is used as a new base sample and sent to the kernelized correlation filter to improve the accuracy of the classifier parameter.An infrared dim and small target tracking algorithm combining kernelized correlation and guided filtering is proposed in this paper.This algorithm can also be applied to track infrared dim and small target under complex sky background,and it has better real-time performance.(4)Most infrared target tracking algorithms cannot be applied to big and small targets at the same time.Based on the particle filter framework,in order to extract the effective features of the infrared target,an joint appearance model combining a sparsity-based generative model and guided filter-based discriminative classifier is used to enhance the performance of the bayesian classifier for effective tracking of infrared targets.(5)Aiming at the problem that the relative distance between the target and the camera cannot be sensed in infrared target tracking.For infrared target distance sensing technology,based on the existing light scattering model in bad weather,a near infrared bispectral imaging system is built using a monochrome camera and two narrow band filters for the first time in this paper.Meanwhile,the visibility and wavelengths are used to estimate the extinction coefficients in the two choosen near infrared wavelengths.And then the proposed physical model of urban building target distance sensing is adopted to perform relative distance sensing.